General materials

3.1.1 Emulsion lipid

Linoleic acid, specified as approximately 60 %, was of the brand Sigma, purchased from Sigma-Aldrich Co., St. Louis, Missouri, USA (catalogue number L1626) and stored in the dark at 5 °C under oxygen-free nitrogen (gas code 152, purchased from BOC New Zealand Ltd., Auckland). According to the supplier, the original plant source was soybean oil and besides linoleic acid the product contained oleic acid and linolenic acids based on gas chromatographic analysis. Several batches were used throughout the project.

3.1.2 Proteins

Whey protein isolate (WPI; product name Alacen 895) and sodium caseinate (NaCas; product name Alanate 180) powders were obtained from Fonterra Cooperative Group New Zealand. Throughout the whole project the same batch of the respective protein type was used. Sodium azide, extra pure, purchased from Merck KGaA, Darmstadt, Germany, was used as microbiological stabiliser in the protein solutions.

3.1.3 Chemicals for lipid hydroperoxide and hexanal determination

Ethanol (95 %), n-hexane (95 %), ethyl acetate (min. 99.5 %), methanol (99.8 %) were obtained from Biolab Ltd., Scoresby, Victoria, Australia, and butan-1-ol from BDH (VWR International, West Chester, PA, USA). Ammonium thiocyanate (min. 98.0 %) and barium chloride (min. 99.0 %) were from Ajax Chemicals (Ajax Finechem Pty.

Ltd., Taren Point, Australia) ferrous sulphate heptahydrate (min. 99.5 %) from Merck and cumene hydroperoxide (88 %) and hexanal (98 %) from Aldrich (Sigma-Aldrich Co., St. Louis, Missouri, USA).

3.2

General methods

3.2.1 Determination of the iron and copper content of the protein powders

Iron and copper contents of the proteins were determined by acid digestion and subsequent ICP-OES (Inductively Coupled Plasma Optical Emission Spectrometry) determination. The measurements were sub-contracted to Hill Laboratories, Hamilton, New Zealand.

3.2.2 Preparation of protein solutions

Protein solutions of various protein contents were prepared with water purified by reverse osmosis (RO). Sodium azide was added to inhibit microbial growth in the protein solutions and emulsions. The concentration of sodium azide in the protein solutions was adjusted to give 200 ppm in the emulsions. Gallaher et al. (2005) had shown that sodium azide had no effect on lipid oxidation in O/W emulsions at this concentration. The solutions were stirred with a magnetic bar for at least 8 h at 50°C, until all protein was dissolved and then stored at 5 ˚C until further use.

3.2.3 Preparation of emulsions

Emulsions containing various protein contents and 10.6 % lipid were prepared with solutions of either WPI or NaCas and linoleic acid. Each solution was added at a temperature of about 5°C to linoleic acid and a coarse emulsion was created by high speed blending at 20500 rpm with a Diax 600 blender from Heidolph Instruments GmbH & Co.KG, Schwabach, Germany. Complete emulsification was achieved with a

two-valve Rannie homogeniser, model APV-1000, from APV New Zealand Ltd., Auckland. The pressure stages generally used for the two valves were 320 bar and 30 bar (320/30 bar), giving physically stable emulsions. The emulsions were passed through the homogeniser three times to ensure complete emulsification.

3.2.4 Acceleration of the lipid oxidation rate by emulsion storage at elevated temperature

The emulsion samples were incubated at 50 ± 0.5 °C in the dark using a Contherm, model 120 MC (Contherm Scientific Ltd., Lower Hutt, New Zealand) as thermostat. These conditions were chosen to allow reasonably fast autoxidation rates of linoleic acid with presumably the same oxidation mechanisms as at room temperature. Other research has found that for incubation tests at 60 °C lipid oxidation correlated well with shelf life tests at room temperature and side reactions were limited compared to higher temperatures (Frankel 1993). Also the results of Hamm et al. (1968) pointed to the same mechanisms in the autoxidation of milk fat up to a temperature of 50 ˚C, according to the volatile compounds produced. The start of the incubation period followed as soon as possible after the emulsification. After emulsion preparation, emulsion aliquots were distributed into containers resulting in either 93 % or 85 % headspace for the lipid hydroperoxide samples or the hexanal samples respectively. The aliquots were 1 ml (lipid hydroperoxide measurements) and 3 ml (hexanal measurements). Sample containers for lipid hydroperoxide measurements were 15 ml plastic centrifuge tubes from Sarstedt (air tight sealed with plastic screw cap) and 20 ml glass vials for gas chromatographic measurements from Phenomenex (air tight sealed with rubber/teflon septum and tin cap). The headspace was regarded as a factor that did not limit lipid hydroperoxide formation as the molar ratio of oxygen initially present in the headspace (20 °C, 1013 hPa) to the highest molarity of lipid hydroperoxides created (WPI (0.5 %) emulsion, d32 = 0.6 µm, in Chapter 4) was 130:1. Samples were incubated in triplicate. Usually, samples were taken at 0, 2 and 4 hours for lipid hydroperoxide and at 0, 12 and 24 h for hexanal measurements. The samples were immediately put into the freezer (-18 °C) at the end of the incubation period. Each container represented an experimental unit for single measurements of either the lipid hydroperoxide or hexanal concentration.

Lipid hydroperoxides and hexanal were determined within 1-2 weeks. In a preliminary experiment samples kept at -18 °C were found to have oxidative stability for at least 3 weeks.

3.2.5 Measurement of the droplet size

Droplet size was measured as Sauter Mean Diameter (d32) in micrometer with a Mastersizer/E from Malvern Instruments Ltd., Malvern, Worcestershire, UK. The presentation code for these measurements was 2NAD. In some experiments, the droplet size was measured as Z-average in micrometer with a Zetasizer Nano-ZS from Malvern Instruments Ltd. In general, the relative refractive index (N) i.e. the ratio of refractive index of fat or lipid globule (1.456) and that of the dispersion medium (1.33) was 1.095. Over the whole incubation period, the droplet size was measured at various times to monitor the physical stability of the emulsions. Independent sample containers were used for measurements at different times. Measurements were carried out in duplicate. This was done separately for all emulsions of an experiment.

3.2.6 Measurement of the pH value

The pH was measured using an Orion 520 pH meter from Thermo Fisher Scientific Inc., Waltham, MA, USA. The pH meter was calibrated daily with a two point calibration at pH 4.0 and 7.0. Emulsions prepared with WPI or sodium caseinate and linoleic acid appeared to be well buffered systems at protein concentrations from 0.5 to 10.0 %. A stable pH of 6.0 ± 0.2 was found over 24 h incubation time (Chapter 4). Therefore the pH was not measured explicitly in all experiments.

3.2.7 Lipid hydroperoxide determination

For the determination of lipid hydroperoxides, a method adapted from Shantha and Decker (1994) and Nuchi et al. (2001) was applied. To frozen 1 ml emulsion aliquots, 5 ml of an effective extraction solvent (ethanol/ethyl acetate/n-hexane, 1:1:1 (V/V/V)),

was added and left thawing at room temperature for 20 min. The solvent mixture had been formerly used successfully by Satue-Gracia et al. (2000) to extract fat from infant formulas. A vortexing procedure to extract the lipids from the samples and centrifugation step followed. Supernatant (0.5 ml) was transferred to a reaction tube with 4 ml of methanol/butanol (2:1 (V/V)). Thiocyanate (30 µl) and ferrous iron solution (30 µl) were added. After each addition, the reaction tubes were vortexed for a few seconds. Three reagent control samples were prepared in the same way, but using the extraction solvent instead of sample extract. The tubes were kept in the dark at room temperature for 20 min before measuring the absorption at 510 nm. A standard curve made with cumene hydroperoxide was used to quantify the lipid hydroperoxides. For this purpose, cumene hydroperoxide was diluted with the extraction solvent to various concentrations. These solutions were used instead of sample extracts in the assay.

Principle of the chemical reaction:

Fe2+ lipid hydroperoxides Fe3+

Fe3+ + thiocyanate coloured complex (absorption maximum at 510 nm)

3.2.8 Hexanal determination

A characteristic secondary oxidation product of linoleic acid is hexanal. The samples were subjected to solid phase microextraction (SPME) that followed hexanal quantification by gas chromatography (GC) using a flame ionisation detector (FID). The principle of SPME is that a solid fibre capable of adsorbing volatiles is exposed to the headspace in the sample vials. After a specified agitation procedure that helps to release volatiles from the sample the fibre is exposed in the injector of the GC and releases the adsorbed compounds into the column gas stream. Characteristics of the SPME procedure were as follows: polydimethylsiloxane fibre (diameter 100 µm), incubation and extraction temperature 50 °C, pre-incubation before extraction at 50 °C for15 min at constant agitation speed, extraction time 20 min at same constant agitation speed. For

the gas chromatograph measurements an Rtx-1701 (30 m, 0.25 mm inner diameter, 1.0 µm film thickness) column was used. A split ratio of 1.0 was applied for the column gas stream. Column temperature program was 45 °C for 1 min, 5 °C/min up to 100 °C and 10 °C/min up to the end temperature of 230 °C. The injector and FID temperature were set at 250 °C. A hexanal external standard curve was created for the quantification of hexanal in the samples. For this purpose hexanal was added at different levels to a freshly prepared emulsion of 4.7 % NaCas and WPI respectively. The average of the two regression curves was used as standard curve.

3.2.9 Spectrophotometric measurements

All spectrophotometric measurements were carried out with an Ultrospec II photometer from Pharmacia LKB (Thermo Fisher Scientific Inc., Waltham, MA, USA). The samples were measured in 4 ml plastic cuvettes.

3.2.10 Statistical analysis

Mean values are shown with the standard error of the mean value (SEM) as error bars to indicate the variability of the single values (calculated in Microsoft Excel). Multifactorial analyses of variance (ANOVA) according to the general linear model were carried out with the lipid hydroperoxide and hexanal concentration as responses. Two sample t-tests were carried out to test if two mean values were significantly different from each other. For these purposes Minitab, version 14 or 15, was used.